mu.micro/mu

Official67 toolsLive

by micro · Go

News, web search, mail, markets, weather, places, files, calendar, contacts. 67 tools, one endpoint.

mu.micro/mu MCP Server

The mu.micro/mu MCP server provides a single endpoint for agent-facing services spanning news, web search, mail, markets, weather, places, files, calendar, and contacts. It exposes 67 tools, and its readme describes an async communication model that routes messaging through one inbox while archiving data locally for search.

🛠️ Key Features

67 tools on one endpoint
Services for news, web search, mail, markets, weather, places, files, calendar, and contacts
Local archiving so archived data is searchable
Async communication via one inbox (email, chat, sms, etc.)

🚀 Use Cases

Using agent tools for retrieving and working with news and web search
Email/SMS/chat-style communication routed through a unified inbox
Agent access to personal/business data such as calendar, contacts, and files

⚡ Developer Benefits

One endpoint that covers multiple service domains
Tool-based integration across many categories (67 tools total)
Locally archived data intended to remain searchable

⚠️ Limitations

The documentation excerpt only states that there are “30+ real world services,” while the server metadata specifies 67 tools; exact service-to-tool mapping is not detailed.

Topics

microagentai

Captured live from the server via tools/list.

agent

Ask the AI agent a question. The agent can search news, markets, web, video, weather, places, and more to answer your question.

Parameters1

prompt

string

required

Your question or request

Raw schema

{
  "type": "object",
  "properties": {
    "prompt": {
      "description": "Your question or request",
      "type": "string"
    }
  },
  "required": [
    "prompt"
  ]
}

apps_build

Build a small app from a natural language description, save it, and return the app details with URL. Apps are one of: a tracker (a list you add entries to, optionally totalling a number), a checklist, or a counter.

Parameters1

prompt

string

required

Description of the app to build (e.g. 'an expense tracker', 'a packing checklist', 'a water intake counter')

Raw schema

{
  "type": "object",
  "properties": {
    "prompt": {
      "description": "Description of the app to build (e.g. 'an expense tracker', 'a packing checklist', 'a water intake counter')",
      "type": "string"
    }
  },
  "required": [
    "prompt"
  ]
}

apps_create

Create a new app — a small, self-contained HTML tool hosted on Mu

Parameters6

description	string	required	Short description of what the app does
html	string	required	The app's HTML content (can include inline CSS and JavaScript, max 256KB)
name	string	required	App name (e.g. Pomodoro Timer)
price	number	optional	Credits charged per use (0 = free, max 1000)
slug	string	required	URL-friendly ID (e.g. pomodoro-timer)
tags	string	optional	Comma-separated tags (optional)

Raw schema

{
  "type": "object",
  "properties": {
    "description": {
      "description": "Short description of what the app does",
      "type": "string"
    },
    "html": {
      "description": "The app's HTML content (can include inline CSS and JavaScript, max 256KB)",
      "type": "string"
    },
    "name": {
      "description": "App name (e.g. Pomodoro Timer)",
      "type": "string"
    },
    "price": {
      "description": "Credits charged per use (0 = free, max 1000)",
      "type": "number"
    },
    "slug": {
      "description": "URL-friendly ID (e.g. pomodoro-timer)",
      "type": "string"
    },
    "tags": {
      "description": "Comma-separated tags (optional)",
      "type": "string"
    }
  },
  "required": [
    "name",
    "slug",
    "description",
    "html"
  ]
}

apps_edit

Edit an existing app you own — update its name, description, tags, icon, HTML code, or price

Parameters7

description	string	optional	New description
html	string	optional	New HTML content (max 256KB)
icon	string	optional	New SVG icon
name	string	optional	New app name
price	number	optional	Credits charged per use (0 = free, max 1000)
slug	string	required	The app's URL slug (e.g. pomodoro-timer)
tags	string	optional	New comma-separated tags

Raw schema

{
  "type": "object",
  "properties": {
    "description": {
      "description": "New description",
      "type": "string"
    },
    "html": {
      "description": "New HTML content (max 256KB)",
      "type": "string"
    },
    "icon": {
      "description": "New SVG icon",
      "type": "string"
    },
    "name": {
      "description": "New app name",
      "type": "string"
    },
    "price": {
      "description": "Credits charged per use (0 = free, max 1000)",
      "type": "number"
    },
    "slug": {
      "description": "The app's URL slug (e.g. pomodoro-timer)",
      "type": "string"
    },
    "tags": {
      "description": "New comma-separated tags",
      "type": "string"
    }
  },
  "required": [
    "slug"
  ]
}

apps_fork

Fork an existing app — creates a copy under your account that you can modify independently

Parameters2

new_slug	string	optional	Slug for the forked copy (optional, auto-generated if empty)
slug	string	required	Slug of the app to fork

Raw schema

{
  "type": "object",
  "properties": {
    "new_slug": {
      "description": "Slug for the forked copy (optional, auto-generated if empty)",
      "type": "string"
    },
    "slug": {
      "description": "Slug of the app to fork",
      "type": "string"
    }
  },
  "required": [
    "slug"
  ]
}

apps_read

Read details of a specific app by its slug

Parameters1

slug

string

required

The app's URL slug (e.g. pomodoro-timer)

Raw schema

{
  "type": "object",
  "properties": {
    "slug": {
      "description": "The app's URL slug (e.g. pomodoro-timer)",
      "type": "string"
    }
  },
  "required": [
    "slug"
  ]
}

apps_run

Run JavaScript code in a sandboxed environment and return the result. Use for calculations, data processing, or any computation the user needs.

Parameters1

code

string

required

JavaScript code to execute. The code runs as a function body — use 'return' to output a value. Has access to mu.ai(), mu.web.fetch(), mu.db and mu.store for platform features.

Raw schema

{
  "type": "object",
  "properties": {
    "code": {
      "description": "JavaScript code to execute. The code runs as a function body — use 'return' to output a value. Has access to mu.ai(), mu.web.fetch(), mu.db and mu.store for platform features.",
      "type": "string"
    }
  },
  "required": [
    "code"
  ]
}

apps_search

Search the apps directory for small, useful tools

Parameters2

query	string	optional	Search query (name, description, or tag)
tag	string	optional	Filter by tag

Raw schema

{
  "type": "object",
  "properties": {
    "query": {
      "description": "Search query (name, description, or tag)",
      "type": "string"
    },
    "tag": {
      "description": "Filter by tag",
      "type": "string"
    }
  }
}

apps_test

Test an app by checking its HTML structure and executing its mu.api calls server-side. Returns which API calls work and which fail.

Parameters1

slug

string

required

The app's URL slug

Raw schema

{
  "type": "object",
  "properties": {
    "slug": {
      "description": "The app's URL slug",
      "type": "string"
    }
  },
  "required": [
    "slug"
  ]
}

block_user

Block a user — hides all their content from your view

Parameters1

user

string

required

User ID to block

Raw schema

{
  "type": "object",
  "properties": {
    "user": {
      "description": "User ID to block",
      "type": "string"
    }
  },
  "required": [
    "user"
  ]
}

blog_create

Create a new blog post

Parameters2

content	string	required	Post content (minimum 50 characters)
title	string	optional	Post title

Raw schema

{
  "type": "object",
  "properties": {
    "content": {
      "description": "Post content (minimum 50 characters)",
      "type": "string"
    },
    "title": {
      "description": "Post title",
      "type": "string"
    }
  },
  "required": [
    "content"
  ]
}

blog_delete

Delete a blog post (author only)

Parameters1

string

required

The blog post ID to delete

Raw schema

{
  "type": "object",
  "properties": {
    "id": {
      "description": "The blog post ID to delete",
      "type": "string"
    }
  },
  "required": [
    "id"
  ]
}

blog_list

Get recent blog posts (titles, snippets and ids; use blog_read for one in full).

Parameters

No parameters.

Raw schema

{
  "type": "object",
  "properties": {}
}

blog_read

Read a specific blog post by ID

Parameters1

string

required

The blog post ID

Raw schema

{
  "type": "object",
  "properties": {
    "id": {
      "description": "The blog post ID",
      "type": "string"
    }
  },
  "required": [
    "id"
  ]
}

blog_update

Update an existing blog post (author only)

Parameters3

content	string	optional	New post content (minimum 50 characters)
id	string	required	The blog post ID to update
title	string	optional	New post title

Raw schema

{
  "type": "object",
  "properties": {
    "content": {
      "description": "New post content (minimum 50 characters)",
      "type": "string"
    },
    "id": {
      "description": "The blog post ID to update",
      "type": "string"
    },
    "title": {
      "description": "New post title",
      "type": "string"
    }
  },
  "required": [
    "id"
  ]
}

chat

Chat with AI assistant

Parameters1

prompt

string

required

The message to send to the AI

Raw schema

{
  "type": "object",
  "properties": {
    "prompt": {
      "description": "The message to send to the AI",
      "type": "string"
    }
  },
  "required": [
    "prompt"
  ]
}

contacts_add

Save someone to your address book. Adding a name already there updates it rather than making a second card.

Parameters4

email	string	optional	Their email address
name	string	required	The person's name
note	string	optional	Anything worth remembering about them
phone	string	optional	Their phone number

Raw schema

{
  "type": "object",
  "properties": {
    "email": {
      "description": "Their email address",
      "type": "string"
    },
    "name": {
      "description": "The person's name",
      "type": "string"
    },
    "note": {
      "description": "Anything worth remembering about them",
      "type": "string"
    },
    "phone": {
      "description": "Their phone number",
      "type": "string"
    }
  },
  "required": [
    "name"
  ]
}

contacts_delete

Remove someone from your address book.

Parameters1

string

required

The contact's id

Raw schema

{
  "type": "object",
  "properties": {
    "id": {
      "description": "The contact's id",
      "type": "string"
    }
  },
  "required": [
    "id"
  ]
}

contacts_find

Look someone up in your address book by name, part of a name, or address. Use this before sending mail to a person named rather than addressed.

Parameters1

query

string

required

A name, part of a name, or an address

Raw schema

{
  "type": "object",
  "properties": {
    "query": {
      "description": "A name, part of a name, or an address",
      "type": "string"
    }
  },
  "required": [
    "query"
  ]
}

contacts_list

List everyone in your address book.

Parameters

No parameters.

Raw schema

{
  "type": "object",
  "properties": {}
}

db_create

Store a record in your database (a named collection). Private by default; set public=true to share it. Pass an id to update a record you own.

Parameters4

collection	string	required	Collection name (e.g. notes, tasks)
data	object	required	The record's fields as a JSON object
id	string	optional	Existing record id to update (optional)
public	boolean	optional	Share the record publicly (default false)

Raw schema

{
  "type": "object",
  "properties": {
    "collection": {
      "description": "Collection name (e.g. notes, tasks)",
      "type": "string"
    },
    "data": {
      "description": "The record's fields as a JSON object",
      "type": "object"
    },
    "id": {
      "description": "Existing record id to update (optional)",
      "type": "string"
    },
    "public": {
      "description": "Share the record publicly (default false)",
      "type": "boolean"
    }
  },
  "required": [
    "collection",
    "data"
  ]
}

db_delete

Delete a record you own by id.

Parameters2

collection	string	required	Collection name
id	string	required	Record id

Raw schema

{
  "type": "object",
  "properties": {
    "collection": {
      "description": "Collection name",
      "type": "string"
    },
    "id": {
      "description": "Record id",
      "type": "string"
    }
  },
  "required": [
    "collection",
    "id"
  ]
}

db_get

Get one record by id from a collection (must be yours, or public).

Parameters2

collection	string	required	Collection name
id	string	required	Record id

Raw schema

{
  "type": "object",
  "properties": {
    "collection": {
      "description": "Collection name",
      "type": "string"
    },
    "id": {
      "description": "Record id",
      "type": "string"
    }
  },
  "required": [
    "collection",
    "id"
  ]
}

db_list

List records in a collection. scope: 'mine' (default), 'public', or 'all' (mine + public). Optional where filter, sort field and limit.

Parameters6

collection	string	required	Collection name
limit	number	optional	Max records (default 50, max 200)
order	string	optional	asc \| desc (default desc)
scope	string	optional	mine \| public \| all (default mine)
sort	string	optional	Data field to sort by
where	object	optional	Filter on data fields, e.g. {"done":false,"priority":{"gte":2}}

Raw schema

{
  "type": "object",
  "properties": {
    "collection": {
      "description": "Collection name",
      "type": "string"
    },
    "limit": {
      "description": "Max records (default 50, max 200)",
      "type": "number"
    },
    "order": {
      "description": "asc | desc (default desc)",
      "type": "string"
    },
    "scope": {
      "description": "mine | public | all (default mine)",
      "type": "string"
    },
    "sort": {
      "description": "Data field to sort by",
      "type": "string"
    },
    "where": {
      "description": "Filter on data fields, e.g. {\"done\":false,\"priority\":{\"gte\":2}}",
      "type": "object"
    }
  },
  "required": [
    "collection"
  ]
}

dismiss

Hide content from your view

Parameters2

id	string	required	Content ID
type	string	required	Content type

Raw schema

{
  "type": "object",
  "properties": {
    "id": {
      "description": "Content ID",
      "type": "string"
    },
    "type": {
      "description": "Content type",
      "type": "string"
    }
  },
  "required": [
    "type",
    "id"
  ]
}

events_create

Schedule a reminder or event. Pass repeat for something recurring, and prompt to have the agent do the work when it fires and mail you the answer — that is how a standing instruction like a morning briefing is set up.

Parameters6

minutes	number	optional	How long it lasts in minutes (default 30). What events_free subtracts.
note	string	optional	Optional extra detail
prompt	string	optional	Optional instruction to run through the agent when it fires, e.g. "brief me on today's news". The answer is mailed to you. This is how you set up something recurring like a morning briefing.
repeat	string	optional	How often it recurs: hourly, daily, weekly or monthly. Omit for once.
title	string	required	What to be reminded about
when	string	required	When to fire, RFC3339 with timezone offset, e.g. 2026-07-22T15:00:00+01:00

Raw schema

{
  "type": "object",
  "properties": {
    "minutes": {
      "description": "How long it lasts in minutes (default 30). What events_free subtracts.",
      "type": "number"
    },
    "note": {
      "description": "Optional extra detail",
      "type": "string"
    },
    "prompt": {
      "description": "Optional instruction to run through the agent when it fires, e.g. \"brief me on today's news\". The answer is mailed to you. This is how you set up something recurring like a morning briefing.",
      "type": "string"
    },
    "repeat": {
      "description": "How often it recurs: hourly, daily, weekly or monthly. Omit for once.",
      "type": "string"
    },
    "title": {
      "description": "What to be reminded about",
      "type": "string"
    },
    "when": {
      "description": "When to fire, RFC3339 with timezone offset, e.g. 2026-07-22T15:00:00+01:00",
      "type": "string"
    }
  },
  "required": [
    "title",
    "when"
  ]
}

events_free

Find when you have nothing booked. Give it how long you need and it returns the open slots, within working hours.

Parameters5

day_end	number	optional	Latest hour to offer, 0-23 (default 18)
day_start	number	optional	Earliest hour to offer, 0-23 (default 9)
from	string	optional	Start of the window, RFC3339 (default now)
minutes	number	optional	How long a slot you need, in minutes (default 30)
to	string	optional	End of the window, RFC3339 (default a week later)

Raw schema

{
  "type": "object",
  "properties": {
    "day_end": {
      "description": "Latest hour to offer, 0-23 (default 18)",
      "type": "number"
    },
    "day_start": {
      "description": "Earliest hour to offer, 0-23 (default 9)",
      "type": "number"
    },
    "from": {
      "description": "Start of the window, RFC3339 (default now)",
      "type": "string"
    },
    "minutes": {
      "description": "How long a slot you need, in minutes (default 30)",
      "type": "number"
    },
    "to": {
      "description": "End of the window, RFC3339 (default a week later)",
      "type": "string"
    }
  }
}

events_list

List your upcoming scheduled events and reminders, soonest first.

Parameters

No parameters.

Raw schema

{
  "type": "object",
  "properties": {}
}

files_delete

Delete a file you own, and its contents.

Parameters1

string

required

The file's id

Raw schema

{
  "type": "object",
  "properties": {
    "id": {
      "description": "The file's id",
      "type": "string"
    }
  },
  "required": [
    "id"
  ]
}

files_get

Read a stored file back by its id. Text comes back as text, anything else as base64.

Parameters1

string

required

The file's id

Raw schema

{
  "type": "object",
  "properties": {
    "id": {
      "description": "The file's id",
      "type": "string"
    }
  },
  "required": [
    "id"
  ]
}

files_list

List your stored files, newest first, with their URLs and ids.

Parameters

No parameters.

Raw schema

{
  "type": "object",
  "properties": {}
}

files_put

Store a file and get a URL for it. Use it to keep something you produced — a report, a CSV, a transcript — and hand back a link.

Parameters4

content	string	required	The file's contents — plain text, or base64 when encoding is base64
encoding	string	optional	"base64" for binary files; omit for text
name	string	required	File name including its extension, e.g. report.csv
type	string	optional	Optional content type, e.g. text/csv. Guessed from the name when omitted

Raw schema

{
  "type": "object",
  "properties": {
    "content": {
      "description": "The file's contents — plain text, or base64 when encoding is base64",
      "type": "string"
    },
    "encoding": {
      "description": "\"base64\" for binary files; omit for text",
      "type": "string"
    },
    "name": {
      "description": "File name including its extension, e.g. report.csv",
      "type": "string"
    },
    "type": {
      "description": "Optional content type, e.g. text/csv. Guessed from the name when omitted",
      "type": "string"
    }
  },
  "required": [
    "name",
    "content"
  ]
}

files_share

Make a stored file readable by anyone holding its URL, or private again.

Parameters2

id	string	required	The file's id
public	boolean	required	True to share, false to make private again

Raw schema

{
  "type": "object",
  "properties": {
    "id": {
      "description": "The file's id",
      "type": "string"
    },
    "public": {
      "description": "True to share, false to make private again",
      "type": "boolean"
    }
  },
  "required": [
    "id",
    "public"
  ]
}

flag

Flag content for moderation

Parameters2

id	string	required	Content ID
type	string	required	Content type (e.g. post, work, app)

Raw schema

{
  "type": "object",
  "properties": {
    "id": {
      "description": "Content ID",
      "type": "string"
    },
    "type": {
      "description": "Content type (e.g. post, work, app)",
      "type": "string"
    }
  },
  "required": [
    "type",
    "id"
  ]
}

hadith

Look up hadith from Sahih Al Bukhari. Pass a book number to get hadiths from that book.

Parameters1

book

number

optional

Book number

Raw schema

{
  "type": "object",
  "properties": {
    "book": {
      "description": "Book number",
      "type": "number"
    }
  }
}

images_generate

Generate an image from a text prompt. Returns a URL to the generated image.

Parameters1

prompt

string

required

Describe the image to generate

Raw schema

{
  "type": "object",
  "properties": {
    "prompt": {
      "description": "Describe the image to generate",
      "type": "string"
    }
  },
  "required": [
    "prompt"
  ]
}

images_search

Search the public image library (community stock) by description. Returns image URLs to reuse.

Parameters1

query

string

required

Search text

Raw schema

{
  "type": "object",
  "properties": {
    "query": {
      "description": "Search text",
      "type": "string"
    }
  },
  "required": [
    "query"
  ]
}

index_search

Search across everything mu knows — indexed news, blog, social and video, plus the user's own mail — and return the most relevant items with ids. Use for 'do you remember', 'what did I get about X', 'search my stuff' and cross-source lookups.

Parameters2

limit	string	optional	Optional max results (default 12)
query	string	required	What to look for

Raw schema

{
  "type": "object",
  "properties": {
    "limit": {
      "description": "Optional max results (default 12)",
      "type": "string"
    },
    "query": {
      "description": "What to look for",
      "type": "string"
    }
  },
  "required": [
    "query"
  ]
}

islam_prayer

Get today's Islamic prayer times (Fajr, Dhuhr, Asr, Maghrib, Isha) for a location, and which prayer is next.

Parameters3

lat	number	required	Latitude of the location
lon	number	required	Longitude of the location
tz	string	optional	IANA timezone of the location, e.g. Europe/London

Raw schema

{
  "type": "object",
  "properties": {
    "lat": {
      "description": "Latitude of the location",
      "type": "number"
    },
    "lon": {
      "description": "Longitude of the location",
      "type": "number"
    },
    "tz": {
      "description": "IANA timezone of the location, e.g. Europe/London",
      "type": "string"
    }
  },
  "required": [
    "lat",
    "lon"
  ]
}

islam_qibla

Get the qibla direction (compass bearing to the Kaaba in Mecca) for a location, and the distance to Mecca.

Parameters2

lat	number	required	Latitude of the location
lon	number	required	Longitude of the location

Raw schema

{
  "type": "object",
  "properties": {
    "lat": {
      "description": "Latitude of the location",
      "type": "number"
    },
    "lon": {
      "description": "Longitude of the location",
      "type": "number"
    }
  },
  "required": [
    "lat",
    "lon"
  ]
}

islam_today

Get today's daily Islamic reminder with verse, hadith, and name of Allah

Parameters

No parameters.

Raw schema

{
  "type": "object",
  "properties": {}
}

mail_address

Get an email address that reaches you. Pass a tag to get your own address (you+tag@) — give that out, then read only its mail with mail_inbox(tag).

Parameters1

tag

string

optional

A label for this address, e.g. "research" or "receipts". Omit for your plain address.

Raw schema

{
  "type": "object",
  "properties": {
    "tag": {
      "description": "A label for this address, e.g. \"research\" or \"receipts\". Omit for your plain address.",
      "type": "string"
    }
  }
}

mail_inbox

Read your mail inbox. Pass a tag to read only mail sent to that plus-address (you+tag@), which is how an agent reads its own mail rather than all of yours.

Parameters1

tag

string

optional

Only mail sent to you+<tag>@ — omit for the whole inbox

Raw schema

{
  "type": "object",
  "properties": {
    "tag": {
      "description": "Only mail sent to you+<tag>@ — omit for the whole inbox",
      "type": "string"
    }
  }
}

mail_send

Send a mail message

Parameters3

body	string	required	Message body
subject	string	required	Message subject
to	string	required	Recipient username or email

Raw schema

{
  "type": "object",
  "properties": {
    "body": {
      "description": "Message body",
      "type": "string"
    },
    "subject": {
      "description": "Message subject",
      "type": "string"
    },
    "to": {
      "description": "Recipient username or email",
      "type": "string"
    }
  },
  "required": [
    "to",
    "subject",
    "body"
  ]
}

markets_list

Get live market prices for cryptocurrencies, stocks, futures, commodities and currencies.

Parameters1

news_list

Get recent news headlines with short summaries balanced across all topics (not dominated by one topic like crypto). Use for general news and briefing requests, then news_read for any article worth expanding.

Parameters2

limit	string	optional	Optional max number of headlines (default 30)
topic	string	optional	Optional topic/category filter (e.g. tech, world, business)

Raw schema

{
  "type": "object",
  "properties": {
    "limit": {
      "description": "Optional max number of headlines (default 30)",
      "type": "string"
    },
    "topic": {
      "description": "Optional topic/category filter (e.g. tech, world, business)",
      "type": "string"
    }
  }
}

news_read

Read one news article in full (title, source, summary and body) by its id from news_headlines, or by article URL.

Parameters1

string

required

Article id (from news_headlines) or article URL

Raw schema

{
  "type": "object",
  "properties": {
    "id": {
      "description": "Article id (from news_headlines) or article URL",
      "type": "string"
    }
  },
  "required": [
    "id"
  ]
}

news_search

Search for news articles

Parameters1

query

string

required

News search query

Raw schema

{
  "type": "object",
  "properties": {
    "query": {
      "description": "News search query",
      "type": "string"
    }
  },
  "required": [
    "query"
  ]
}

places_eta

How long it takes to travel between two places, by road rather than as the crow flies. Use it to answer whether somewhere is worth going to, or when to leave.

Parameters7

from	string	optional	Where the journey starts, e.g. "King's Cross, London"
from_lat	number	optional	Start latitude, if already known
from_lon	number	optional	Start longitude, if already known
mode	string	optional	How to travel: drive (default), walk, cycle or transit
to	string	optional	Where the journey ends, e.g. "Heathrow Airport"
to_lat	number	optional	End latitude, if already known
to_lon	number	optional	End longitude, if already known

Raw schema

{
  "type": "object",
  "properties": {
    "from": {
      "description": "Where the journey starts, e.g. \"King's Cross, London\"",
      "type": "string"
    },
    "from_lat": {
      "description": "Start latitude, if already known",
      "type": "number"
    },
    "from_lon": {
      "description": "Start longitude, if already known",
      "type": "number"
    },
    "mode": {
      "description": "How to travel: drive (default), walk, cycle or transit",
      "type": "string"
    },
    "to": {
      "description": "Where the journey ends, e.g. \"Heathrow Airport\"",
      "type": "string"
    },
    "to_lat": {
      "description": "End latitude, if already known",
      "type": "number"
    },
    "to_lon": {
      "description": "End longitude, if already known",
      "type": "number"
    }
  }
}

places_nearby

Find all places of interest near a given location

Parameters4

address	string	optional	Address or postcode to search near
lat	number	optional	Latitude of the search location
lon	number	optional	Longitude of the search location
radius	number	optional	Search radius in metres, 100–5000 (default 500)

Raw schema

{
  "type": "object",
  "properties": {
    "address": {
      "description": "Address or postcode to search near",
      "type": "string"
    },
    "lat": {
      "description": "Latitude of the search location",
      "type": "number"
    },
    "lon": {
      "description": "Longitude of the search location",
      "type": "number"
    },
    "radius": {
      "description": "Search radius in metres, 100–5000 (default 500)",
      "type": "number"
    }
  }
}

places_search

Search for places by name or category, optionally near a location

Parameters5

near	string	optional	Location name or address to search near
near_lat	number	optional	Latitude of the search location
near_lon	number	optional	Longitude of the search location
query	string	required	Search query (e.g. cafe, pharmacy, Boots)
radius	number	optional	Search radius in metres, 100–5000 (default 1000)

Raw schema

{
  "type": "object",
  "properties": {
    "near": {
      "description": "Location name or address to search near",
      "type": "string"
    },
    "near_lat": {
      "description": "Latitude of the search location",
      "type": "number"
    },
    "near_lon": {
      "description": "Longitude of the search location",
      "type": "number"
    },
    "query": {
      "description": "Search query (e.g. cafe, pharmacy, Boots)",
      "type": "string"
    },
    "radius": {
      "description": "Search radius in metres, 100–5000 (default 1000)",
      "type": "number"
    }
  },
  "required": [
    "query"
  ]
}

quran

Look up a Quran chapter or verse. Pass chapter number (1-114) and optionally a verse number.

Parameters2

chapter	number	required	Chapter number (1-114)
verse	number	optional	Verse number (optional, returns full chapter if omitted)

Raw schema

{
  "type": "object",
  "properties": {
    "chapter": {
      "description": "Chapter number (1-114)",
      "type": "number"
    },
    "verse": {
      "description": "Verse number (optional, returns full chapter if omitted)",
      "type": "number"
    }
  },
  "required": [
    "chapter"
  ]
}

quran_search

Search the Quran, Hadith, and names of Allah using semantic search. Ask a question in natural language.

Parameters1

query

string

required

Question or search query

Raw schema

{
  "type": "object",
  "properties": {
    "query": {
      "description": "Question or search query",
      "type": "string"
    }
  },
  "required": [
    "query"
  ]
}

save

Bookmark content for later

Parameters2

id	string	required	Content ID
type	string	required	Content type (e.g. post, work, app)

Raw schema

{
  "type": "object",
  "properties": {
    "id": {
      "description": "Content ID",
      "type": "string"
    },
    "type": {
      "description": "Content type (e.g. post, work, app)",
      "type": "string"
    }
  },
  "required": [
    "type",
    "id"
  ]
}

saved_list

List the items you've saved for later (bookmarks), with their links.

Parameters

No parameters.

Raw schema

{
  "type": "object",
  "properties": {}
}

social_list

Get the latest social posts from the network.

Parameters

No parameters.

Raw schema

{
  "type": "object",
  "properties": {}
}

social_search

Search social media posts

Parameters1

query

string

required

Search query for social posts

Raw schema

{
  "type": "object",
  "properties": {
    "query": {
      "description": "Search query for social posts",
      "type": "string"
    }
  },
  "required": [
    "query"
  ]
}

stream_list

Read the platform event stream — user messages, agent responses, system events (markets, news, reminders)

Parameters

No parameters.

Raw schema

{
  "type": "object",
  "properties": {}
}

stream_post

Post a message to the stream. Mention @micro to get an AI response.

Parameters1

content

string

required

Message text (max 1024 chars). Use @micro to invoke the AI agent.

Raw schema

{
  "type": "object",
  "properties": {
    "content": {
      "description": "Message text (max 1024 chars). Use @micro to invoke the AI agent.",
      "type": "string"
    }
  },
  "required": [
    "content"
  ]
}

unblock_user

Unblock a previously blocked user

Parameters1

user

string

required

User ID to unblock

Raw schema

{
  "type": "object",
  "properties": {
    "user": {
      "description": "User ID to unblock",
      "type": "string"
    }
  },
  "required": [
    "user"
  ]
}

unsave

Remove a saved bookmark

Parameters2

id	string	required	Content ID
type	string	required	Content type

Raw schema

{
  "type": "object",
  "properties": {
    "id": {
      "description": "Content ID",
      "type": "string"
    },
    "type": {
      "description": "Content type",
      "type": "string"
    }
  },
  "required": [
    "type",
    "id"
  ]
}

video_list

Get the latest videos from curated channels.

Parameters

No parameters.

Raw schema

{
  "type": "object",
  "properties": {}
}

video_search

Search for videos

Parameters1

query

string

required

Video search query

Raw schema

{
  "type": "object",
  "properties": {
    "query": {
      "description": "Video search query",
      "type": "string"
    }
  },
  "required": [
    "query"
  ]
}

wallet_balance

Get your balance: credits (what calls are charged in), plus your Base address and USDC balance for topping up.

Parameters

No parameters.

Raw schema

{
  "type": "object",
  "properties": {}
}

weather_forecast

Get the weather forecast for a location (current conditions plus the next few days).

Parameters2

lat	number	required	Latitude of the location
lon	number	required	Longitude of the location

Raw schema

{
  "type": "object",
  "properties": {
    "lat": {
      "description": "Latitude of the location",
      "type": "number"
    },
    "lon": {
      "description": "Longitude of the location",
      "type": "number"
    }
  },
  "required": [
    "lat",
    "lon"
  ]
}

web_fetch

Fetch a web page and return its cleaned readable content (strips ads, popups, navigation)

Parameters1

url

string

required

The URL to fetch

Raw schema

{
  "type": "object",
  "properties": {
    "url": {
      "description": "The URL to fetch",
      "type": "string"
    }
  },
  "required": [
    "url"
  ]
}

web_search

Search the web for current information and news

Parameters1

query

string

required

Search query

Raw schema

{
  "type": "object",
  "properties": {
    "query": {
      "description": "Search query",
      "type": "string"
    }
  },
  "required": [
    "query"
  ]
}

README.md11 KBView on GitHub

Mu

The runtime for Micro, a personal assistant.

Overview

Ask on the web, send an email, or message it from your phone. Mu runs the assistant, its tools, and your saved conversations in one Go binary that you can host yourself. Try the hosted instance at micro.mu.

Features

The web starts with one input. Ask a question or give an instruction; the answer appears below it. The prompt moves up on the first request and stays there while you read and continue the conversation.

Inbox keeps saved conversations and incoming messages together. Return to a web conversation and continue where you left off.
Account holds your profile, connections, client credentials, and billing.
Admin, visible to administrators, holds users, settings, logs, and server controls.

One shared stylesheet (/mu.css) and browser script (/mu.js) serve the pages. HTML is rendered in Go. There is no frontend build step. The site includes a manifest and service worker so it can be installed as a PWA.

Clients

Open Contact in the footer, or Account → Reach Micro, for the addresses and numbers configured on your instance. Add Micro to your phone's contacts from that page.

Channel	How it fits
Web	Start a conversation or reopen one from Inbox.
Email	Send or forward a message to `agent@your-domain` from your verified email address. Include what you want the assistant to do.
SMS	Verify your phone number in Account, then text the configured number.
WhatsApp	Use your verified number to message the configured WhatsApp sender.
XMPP	Connect with your account and a Chat token from Client access, then message the agent.

Email, SMS, WhatsApp, and XMPP require the corresponding server configuration. SMS and WhatsApp use Twilio; they are not enabled merely by installing Mu. Contact lists the configured ways to reach the assistant. Incoming mail must pass the sender checks; account identity comes from verified addresses and numbers, not from whatever a message claims.

Conversations from the different channels appear in one inbox, but are still separate threads. Switching from a web conversation to a fresh SMS does not automatically continue that exact thread. Replies retain their channel context. Forwarding mail to your own mailbox stores it; addressing an agent asks it to act. WhatsApp replies are subject to the provider's messaging window.

Agents

Micro is the default agent. Agents have instructions and a permitted set of tools. Services provide those tools: mail, files, calendar, search, weather, notes, shell, and more. You ask for an outcome; the agent chooses the tools.

Google

Optional Google connections provide access to Gmail, Calendar, Contacts, and Drive with your consent. User-created agents can have different instructions and access; mail to you+research@your-domain addresses your Research agent.

Google reads use the signed-in account's connection and granted scope. Gmail searches default to the last 30 days unless an explicit date range is supplied. The assistant carries at most 24 recent messages within a 16,000-character history budget. Older conversations and saved notes are read through permitted tools when needed, rather than automatically added to every question. Disconnecting Google stops new reads; it does not erase answers already saved in your Inbox.

Work

An explicitly requested job can run in the background and return its result to the originating conversation. Execution lives under agent/work; task records live in service/tasks. The personal daily brief remains available. Other unsolicited model-generated feeds are disabled.

Protocols

Account → Client access shows connection details and creates tokens with an explicit choice of Mail, Chat, or both.

IMAP reads the inbox; SMTP submission sends mail using a Mail token.
XMPP uses a Chat token.
SFTP transfers stored files using a registered SSH key.
SSH opens an interactive terminal in your account's sandbox using that key. It does not grant access to the host machine. Remote exec and port forwarding are not supported.

SSH/SFTP must be enabled by the operator. Legacy SHELL_SHARED execution is blocked; existing shared workspaces need migration to per-account containers. Mail and XMPP public TLS endpoints require the proxy setup described in the installation guide. The page reports configured settings; it cannot verify that external DNS, ports, or proxies work.

Install

bash

curl -fsSL https://raw.githubusercontent.com/micro/mu/main/install.sh | sh
mu setup
mu --serve

Open http://localhost:8080. Initial administrator setup depends on the instance's bootstrap configuration; see the installation guide.

Configure an AI provider with mu setup. Supported settings include ANTHROPIC_API_KEY, ATLASCLOUD_API_KEY, GEMINI_API_KEY, OPENROUTER_API_KEY, or an OpenAI-compatible OPENAI_BASE_URL. Individual tools may need provider keys, such as BRAVE_API_KEY for web search or YOUTUBE_API_KEY for video. Google sign-in and payments are optional.

From source:

bash

git clone https://github.com/micro/mu
cd mu
go install
mu setup
mu --serve

Or run docker compose up from the checkout. See the installation guide for domains, TLS, mail, messaging, sandbox configuration, and deployment.

CLI

The binary also acts as a client. It defaults to the hosted instance; set MU_URL or use mu login https://your.host for your own server.

bash

mu ask "What needs my attention?"
mu inbox list
mu help

CLI and API operations require a credential with the appropriate API or service permissions. Client access offers separate Mail/Chat protocol tokens, product tokens with selected capabilities and optional actions, and tokens restricted to selected services. Protocol tokens do not grant CLI, agent API, or MCP access. Agent API access can execute your account’s agents with their configured tools; use service scopes when a client should reach only specific capabilities.

API

Service operations remain at /api/v1/<service>/<method> and /mcp, documented at /api and /tools. The x402 host exposes the same service contract with payment handling. Credentials restrict access; they do not switch catalogues.

Product clients use the existing resource URLs: /agent, /agent/<name>, /inbox and /work. Content-Type selects JSON or form input; Accept selects JSON or HTML output. /developers documents the supported requests. There are no separate product API or MCP endpoints.

bash

curl "$MU_URL/agent" \
  -H "Authorization: Bearer $MU_TOKEN" \
  -H 'Content-Type: application/json' \
  -d '{"prompt":"What needs my attention?"}'

The response includes text and thread; send the thread identifier back to continue. Requests and credentials belong in request bodies and headers. A separately configured x402 host provides paid service calls.

Development

bash

go build ./...

Pages, styles, scripts, and bundled service data are embedded in the binary; changes require rebuilding. Server settings are available in Admin → Settings.

VERSION names the release. Updating it on main runs the release workflow, which builds Linux and macOS binaries, creates the version tag, and publishes the release. A version-tag push or manual workflow dispatch is also supported.

License

AGPL 3.0

App builds are queued and return an ID immediately. Use Apps.BuildStatus with that ID, Apps.Read with the returned slug, or the build status URL to follow the result. Reuse request_key when retrying a submission; without one, identical requests by the same account resolve to the same build. Build records and model output are checkpointed under the data directory. Recovery retains the same ID, counts interrupted model calls against the three-attempt limit, and never restarts a completed model call just to retry saving. Generated apps start private. An interrupted provider request can still incur a charge. This queue assumes one server process owns the data directory; it is not a distributed worker queue.

Anonymous model prompts are disabled by default. ALLOW_GUEST_AI=true explicitly re-enables them with guest limits: 10 calls per browser, 30 per IP and 100 across the instance per hour (GUEST_MAX_PER_CLIENT, GUEST_MAX_PER_IP, GUEST_MAX_TOTAL, GUEST_WINDOW_MINUTES). These count admitted guest requests, not individual model/tool calls, and reset on restart. Public pages remain readable; these controls do not prevent all scraping. Configure TRUSTED_PROXY correctly so client addresses are resolved at the intended boundary.

Rate limits

Dynamic HTTP routes are limited before page handling. Public reading stays open; priced service operations require an authenticated, verified/approved account or an authenticated x402 wallet using the existing payment gate. Assistant runs have separate account budgets, including failed attempts. Account and operator billing exemptions do not bypass these attempt budgets.

Setting	Default
`HTTP_MAX_PER_MINUTE`	300 per IP, including authenticated traffic
`HTTP_GUEST_MAX_PER_MINUTE`	60 per IP
`HTTP_GUEST_MAX_PER_HOUR`	300 per IP
`AUTH_MAX_PER_15_MINUTES`	20 non-GET authentication requests per IP
`ASSISTANT_MAX_PER_HOUR`	60 runs per account
`ASSISTANT_MAX_PER_DAY`	300 runs per account
`ASSISTANT_MAX_CONCURRENT`	2 runs per account
`PAID_MAX_PER_HOUR`	300 priced service attempts per account
`PAID_MAX_PER_DAY`	1,000 priced service attempts per account

These are fixed windows beginning with the first attempt, not monetary spending caps. Positive settings override defaults; zero does not disable protection. Known /mu.css, /mu.js, manifest, favicon, robots and /static/ assets bypass HTTP limits. IPv6 clients share a /64 limit. Configure TRUSTED_PROXY for your actual reverse proxies; forwarded chains are read from the trusted end.

Attempt counters are committed to abuse.db before work starts. They survive a process restart when the same data directory is retained. Expired rows are pruned; the store caps identities at 20,000 and refuses new ones when full. Storage failure refuses protected requests. Daily successful-operation counts are also persisted. This supports one server process per data directory, not multiple independent replicas. Existing credit billing remains separate.

HTTP limits return 429 and Retry-After; protection-storage or admission-queue failures return 503 and Retry-After. Tool protocol errors retain their native format. /admin/traffic?window=day includes a 24-hour view, HTTP status/endpoint counts and instrumented model call counts. Requests rejected before identity validation appear under unattributed and http-refused, not under a guessed account. New counters start at deployment; 2xx HTTP responses can still contain application or MCP errors. Agent model counts are reported when the run records its usage; an interrupted run may not report them. These application limits bound ordinary abuse, not distributed volumetric attacks or all copying of public content.

Install

Tools (67)

agentapps_buildapps_createapps_editapps_forkapps_readapps_runapps_search

Remote endpoint

Streamable HTTP

Hosted server - connect over the network, no local install.

url

https://micro.mu/mcp

Add to VS Code Add to Cursor

claude_desktop_config.json

json

{
  "mcpServers": {
    "mu": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://micro.mu/mcp"
      ]
    }
  }
}

Desktop config is stdio-only; this bridges via mcp-remote. Native remote: Settings > Connectors.